The deserts of southern Asia have high potential for solar energy production due to low latitude and nearly cloudless summers. This study uses satellite data from the MIDAS climatology and cloud data from CM SAF to examine the effects of aerosols, dust, and clouds on surface solar radiation (SSRD) and direct normal irradiance (DNI) from 2004 to 2023. It also evaluates their impact on solar energy facilities. Results show that dust aerosol concentration (DUAOD550) and total aerosols (AOD550) have a weak positive correlation with SSRD and DNI (R² between 0.02 and 0.048). From 2016 to 2024, the average dust optical depth (Dust AOD) decreased in countries like Pakistan (from 0.53 to 0.11) and Afghanistan (from 0.25 to 0.1). Global horizontal irradiance (GHI) and direct normal irradiance (DNI) increased in some areas, such as the Paracel Islands, but decreased in Pakistan and Aksai Chin. These changes affect the efficiency of solar energy systems and require consideration in facility design.
This research details the development of a radar system for vehicle speed detection and license plate recognition, integrating an IR sensor, ESP32 microcontroller, Arduino Uno, and MATLAB software. The system’s primary objective is to accurately measure vehicle speeds and extract number plates using a combination of hardware and software components. The hardware setup includes an IR sensor for vehicle detection and an ESP32 with a Wi-Fi module for data transmission, while an Arduino Uno interfaces these components and collects sensor data. Software aspects involve Arduino IDE for programming the Uno and MATLAB for advanced image processing and analysis, facilitating efficient extraction of number plates. Optical Character Recognition (OCR) algorithms are then employed to recognize alphanumeric characters. By analyzing sequential images captured at specific intervals, the system estimates vehicle speed. This automated system has potential applications in traffic management, law enforcement, and parking enforcement, offering a comprehensive solution for monitoring traffic, detecting license plates, and estimating vehicle speed.This research directly supports SDG 9 (Industry, Innovation and Infrastructure) by creating innovative technological infrastructure, and also contributes to SDG 11 (Sustainable Cities and Communities) by enhancing road safety and traffic management.
In recent years, China has made significant progress in the field of satellite computing and is building an integrated space-based intelligent infrastructure that unifies sensing and computing in space. These advancements are first reflected in the rapid growth of the number of satellites. Secondly, multi-satellite constellations enable on-orbit edge computing, inter-satellite collaboration, and AI inference, driving satellite systems to evolve from data acquisition to intelligent processing. China possesses both land-based and sea-based launch platforms, demonstrating a steadily maturing all-domain space launch capability. Satellite intelligence has promoted technological progress in many application areas, such as ecological environment management and biodiversity conservation. In this talk some representative cases are to be introduced.
In this talk, we will explore the potential implementation of several physical layer blocks using AI/ML algorithms, including IF estimation, symbol rate estimation, channel estimation, modulation classification, early ACK/NACK activation, and AI/ML-based LDPC decoding. We will conclude with a discussion of possible future directions for AI/ML in wireless communications
The growing demand for halal‐certified products underscores the need to understand the key determinants influencing halal purchase decisions. This study examines the structural relationships between halal awareness, halal certification, attitude, religiosity, and purchase intention. Using a quantitative approach, the model reveals that halal certification is the strongest predictor of purchase intention (β = 0.516), indicating that consumers rely heavily on verified certification when evaluating product compliance. Halal awareness positively influences attitude (β = 0.319), although its direct effect on purchase intention remains weak (β = 0.070), suggesting that knowledge alone does not translate into behavioral intention without credible assurance. Attitude contributes significantly to purchase intention (β = 0.237), confirming its mediating role in halal decision-making. Religiosity demonstrates a minimal direct effect (β = 0.089); however, its interaction with halal certification strengthens purchase intention (β = 0.263), while its interaction with awareness negatively moderates intention (β = –0.214). These findings suggest that halal consumer behavior is shaped by institutional trust and perceived product legitimacy rather than religiosity or awareness alone. The study offers practical implications for halal product regulators, marketers, and certification agencies seeking to enhance consumer compliance and market acceptance
The exponential growth of digital communication, mobile devices, cloud platforms, and Internet of Things (IoT) ecosystems has intensified the energy footprint of information networks, raising urgent concerns regarding environmental sustainability and long-term viability. Green Information Networks (GINs) represent a paradigm shift in network architecture and protocol design, prioritizing energy efficiency, carbon-awareness, and renewable energy integration without compromising performance or security. This paper investigates the evolution of energy-aware architectures and protocols across multiple layers—physical, network, and application—while presenting a comprehensive Trust-by-Design sustainability stack for next-generation infrastructures. Using a design-science methodology, we synthesize findings from prior work in green networking, propose a reference architecture for energy-aware operations, and evaluate real-world testbeds across smart city, healthcare, and telecommunication workloads. Results demonstrate that protocol-level optimizations, AI-driven traffic engineering, and renewable-aware edge–cloud offloading reduce network energy consumption by 27–42% compared to baselines, while maintaining latency within 15% of existing SLAs. We show that carbon intensity per gigabit transferred can be reduced to below 30 gCO₂e in optimized deployments, aligning with 2030 Net Zero trajectories. Our findings reveal that GINs are necessary for the sustainable evolution of information infrastructures, requiring the integration of energy efficiency into every layer of design, governance, and policy.
Wireless Mesh Networks (WMNs) represent a class of decentralized, self-organizing wireless systems widely adopted for real-time communication and broadband access in heterogeneous environments. In WMNs, multiple mesh routers and client nodes collaborate to relay data across multi-hop links through wireless access points (WAPs). The distributed nature of WMNs enables robust connectivity without relying on centralized infrastructure, making them suitable for applications such as community broadband, disaster recovery, and industrial IoT. Despite their flexibility, WMNs suffer from significant latency and packet loss due to dynamic topologies, node mobility, and multi-hop data forwarding. As the number of nodes increases within a given coverage region, data hopping and link disruptions lead to degraded Quality of Service (QoS), particularly in Voice over Internet Protocol (VoIP) applications where delay and jitter directly affect speech quality. In such cases, if an intermediate node fails, data packets must reroute through alternative hops, increasing end-to-end delay and reducing packet delivery performance. To address these challenges, this study proposes a Hybrid Routing Protocol (HRP) designed to enhance VoIP performance in WMNs by integrating intelligent node selection and adaptive path optimization. The HRP operates through three major phases: (1) Active Node Selection (ANS) for optimized node deployment and coverage control; (2) Hopfield Neural Network (HNN) for predictive route selection based on learned network dynamics; and (3) Particle Swarm Optimization (PSO) for dynamic path recovery under uncertain link conditions. The proposed HRP is simulated under varying node densities and distances to evaluate its adaptability and reliability. Comparative performance analysis against conventional routing protocols demonstrates that HRP achieves superior packet delivery ratio (PDR), lower end-to-end delay, and higher throughput, making it an efficient routing solution for real-time VoIP applications over Wireless Mesh Networks.
In our earth, less than One-third (29%) of its surface is covered by land and the remaining more than two-third (71%) remains water bodies like ocean, river, lake, etc. The oceans are rich in almost all resources that are available in land. Hence, there is a need to discover unexplored underwater assets, treasures, minerals etc. In the past recent years, the development in underwater technology and understanding of the ocean by human has made many researchers and industries to look into exploration of the ocean. Underwater exploration also helps to realize the environmental degradation in how humans are affecting and being exaggerated by changes in the ocean environment. The underwater has many applications in the field of defense, survey, archaeology, mines, navigation, animal bio acoustics etc. To mention a few which has more impact on the society are tsunami buoys, desalination system, protection of the coastal region etc. All the application involves various types of underwater sensors. This talk is going to concentrate on the fundamental to advanced underwater sensors that are used for different applications from shallow water to deep water. This presentation also deals with real time data collection using important underwater sensors and its post processing for accurate detection of the required application. The talk will touch on the hydrophones that are used for ambient noise data collection and its effect on the signal transmission in underwater channel. The Grab and Water sample Collector for Geo acoustic Inversion Study, SVP & CTD for Underwater acoustic signal analysis, an unique Underwater Battery that uses Ocean water as electrolyte, Side Scan Sonar (SSS) to obtain high-resolution seabed images, Multibeam Echo Sounder (MBES) for Seafloor mapping, Optical Cameras and Remotely Operated Vehicles (ROV) to acquire data for image processing, and Permanent Magnetic Linear Generator (PMLG) to harvest energy from sea wave. Autonomous Underwater Vehicles (AUVs) and remotely operated vehicles (ROVs) are extensively used in the scrutiny of submarines, pipelines, seabed mapping, and underwater image analysis. The acoustic signal transmission in underwater is strongly affected by the interaction of sea-bottom sediments. The geo-properties like mean grain size, porosity, grain density of marine sediments affect the sound propagation. These properties vary from location to location; this is dependent on depositional characteristics of the particular location. The above are some of the underwater sensors that will be discussed in the talk along with its application and impact on underwater research.
The rise of new-generation mobile networks, including 5G and the impending 6G, poses formidable technical hurdles in attaining the ambitious benchmarks set by the research and industry communities. These challenges encompass accommodating a multitude of devices on a single network, ensuring ultra-reliable low-latency communication, sustaining adaptability and dynamism, and delivering ample high-quality bandwidth. To tackle these complexities effectively, there is an escalating demand for a unified approach amalgamating network management and control, featuring autonomous and adaptable actions. The presented Distributed Artificial Intelligence (DAI) framework harnesses Belief Desire Intention (BDI) agents endowed with machine learning capabilities, denoted as BDIx agents, because it uses ML under believes. These agents are dispersed across mobile devices, forming a multi-agent system (MAS) that incorporates Fuzzy Logic and Back-Propagation Neural Networks for Reinforcement Learning at the agents' perceptual and cognitive tiers. A prime illustration of the DAI framework is demonstrated in the context of Device-to-Device (D2D) communication within 5G and beyond networks. D2D communication's decentralized nature, coupled with a multitude of user devices (User Equipment or UEs), presents an ideal platform to showcase the capabilities of the DAI framework. By integrating BDIx agents into D2D UEs, it becomes possible to circumvent the conventional Base Station (BS) and establish direct links among neighboring UEs. This approach promises enhancements in spectral and energy efficiency, data rates, throughput, latency, interference management, and fairness. Given the manifold challenges introduced by D2D communication in 5G and 6G networks, the DAI framework is anticipated to play a pivotal role in surmounting these obstacles and fostering innovations in Artificial Neural Networks and other facets of these dynamic mobile networks. In my presentation, I will showcase the ADROIT6G EU project, which utilizes the proposed framework to realize BDIx agents. ADROIT6G's primary goal is to establish innovative research principles to advance low Technology Readiness Level technologies in preparation for the future 6G network architectures. This project seeks to enhance the current service_x0002_based structures of 5G mobile networks by developing and validating a forward-looking, cognitive 6G architecture. This will be achieved through a fully distributed paradigm driven by Artificial Intelligence, deploying functional elements as virtual functions in cloud-native environments spanning the far-edge, edge, and cloud domains, and involving multiple stakeholders. These advancements aim to deliver improved performance, greater control, increased transparency in digital service interactions, support for innovative applications, and societal acceptance, marking a significant step towards the evolution of next-generation 6G networks
Research in the telecommunication domain is running at unprecedented pace in several areas spanning from propagation to multimedia processing and distribution, from optical to non-terrestrial telecommunication networks, from network intelligence for resource management to machine learning for security. Therefore, it is expected that several breakthroughs will stem in different research areas. However, the actual achievement of the desired impact in terms of new market opportunities and increased competitiveness of the relevant industries of such breakthroughs will happen only if they are inset into a coherent architectural framework. Objective of this talk is to discuss the main features of an open architecture capable of integrating innovative solutions developed in different communication and networking domains into a holistic framework that is coherent from both a technical and a business point of view. Such architecture should encompass all elements of the future network with special emphasis on the network edge and extreme edge where network functions will be executed. To achieve such goals the network architecture must be based on the concept of Digital Twin, suitably extended to meet the needs of a future telecommunication network, capable of supporting 6G and beyond services and considering sustainability as a major design driving force and thus focusing on key value indicators (KVIs) besides the key performance indicators (KPIs). In the envisioned architecture DTs will be implemented as a composition of microservices running across a fully distributed peer-to-peer platform to guarantee robustness, resilience and to remain open to the entrance of new players. Also, we will discuss how such network architecture can exploit generative artificial intelligence to make continuous innovation possible.
Efficient resource allocation in Device-to-Device (D2D) communication within 6G networks is crucial for enhancing overall network performance and efficiency. This paper presents a novel Deep Learning (DL) based approach for Radio Resource Allocation (RRA), leveraging Distributed Artificial Intelligence (DAI) using Belief-Desire-Intention eXtended (BDIx) agents, dynamic feedback allocation, and a Deep Feedback Neural Network (DFBNN). Additionally, Federated Learning (FL) is integrated to enable distributed training across BDIx agents, serving as D2D Relays (D2DR) or D2D Multihop Relays (D2DMHR), ensuring data privacy and reducing communication overhead. The proposed method is thoroughly evaluated against traditional graph-based and game-theoretic algorithms and Deep Feedforward Neural Networks (DFNN). Results demonstrate significant improvements in interference management, data rate, and execution time. By providing scalable, adaptive, and resilient resource allocation, this proposed method meets the stringent requirements of 6G applications, paving the way for more efficient and reliable network operations.
<strong><em>The augmentation of fake news across online platforms has come forth as a challenge to society and threat to democracy. Fake news gnaws confidence in reliable news sources and threatens social cohesion and belief in democracy. Fake news comes from different sources and spreads like a wildfire. It becomes difficult to distinguish the authenticity of real news from fake news. While numerous studies have addressed fake news detection using machine learning algorithms, many conventional approaches are limited by their reliance on manual feature engineering or an incomplete understanding of linguistic context. This paper works on a more advanced approach using a fine-tuned Bidirectional Encoder Representations from Transformers (BERT) model to overcome this limitation. This research work puts emphasis on fine-tuning a pre-trained BERT model on a task-specific news dataset. Fine tuning can significantly improve detection accuracy. An extensive study has been carried out on the ISOT dataset taken from the University of Victoria that consists of thousands of real and fake news articles. The model used in the research achieved an accuracy of 99.97%, precision of 100%, F-1 score of 99.97% and recall of 99.94%, validating its superiority over previously reported methods. </em></strong>
